KRAP regulates mitochondrial Ca²⁺ uptake by licensing IP₃ receptor activity and stabilizing ER–mitochondrial junctions

ABSTRACT

Inositol 1,4,5-trisphosphate (IP₃) receptors (IP₃Rs) are high-conductance channels that allow the regulated redistribution of Ca²⁺ from the endoplasmic reticulum (ER) to the cytosol and, at specialized membrane contact sites (MCSs), to other organelles. Only a subset of IP₃Rs release Ca²⁺ to the cytosol in response to IP₃. These ‘licensed’ IP₃Rs are associated with Kras-induced actin-interacting protein (KRAP, also known as ITPRID2) beneath the plasma membrane. It is unclear whether KRAP regulates IP₃Rs at MCSs. We show, using simultaneous measurements of Ca²⁺ concentration in the cytosol and mitochondrial matrix, that KRAP also licenses IP₃Rs to release Ca²⁺ to mitochondria. Loss of KRAP abolishes cytosolic and mitochondrial Ca²⁺ signals evoked by stimulation of IP₃Rs via endogenous receptors. KRAP is located at ER–mitochondrial membrane contact sites (ERMCSs) populated by IP₃R clusters. Using a proximity ligation assay between IP₃R and voltage-dependent anion channel 1 (VDAC1), we show that loss of KRAP reduces the number of ERMCSs. We conclude that KRAP regulates Ca²⁺ transfer from IP₃Rs to mitochondria by both licensing IP₃R activity and stabilizing ERMCSs.

Summary: KRAP regulates transfer of Ca²⁺ from IP₃ receptors to mitochondria by both stabilizing membrane contact sites between the ER and mitochondria and licensing IP₃ receptors to respond to IP₃ at these junctions.

INTRODUCTION

Inositol 1,4,5-trisphosphate (IP₃) receptors (IP₃Rs) are Ca²⁺ channels expressed in the membranes of the endoplasmic reticulum (ER), where their opening, after binding of IP₃ and Ca²⁺, allows Ca²⁺ to leak rapidly from the ER (Foskett et al., 2007; Prole and Taylor, 2019b). The Ca²⁺ released can then diffuse into the cytosol to increase the cytosolic Ca²⁺ concentration, which can then regulate diverse cellular activities. Ca²⁺ can also be delivered to other membranes to which the ER is closely apposed at membrane contact sites (MCSs) (Helle et al., 2013; Roest et al., 2017). Opening of IP₃Rs within MCSs between the ER and lysosomes or mitochondria, for example, delivers Ca²⁺ locally at concentrations sufficient to fuel uptake by the low-affinity Ca²⁺ uptake systems of these organelles (De Stefani et al., 2011; Atakpa et al., 2018). The most thoroughly studied of these exchanges occurs within ER–mitochondrial MCSs (ERMCSs), where IP₃Rs deliver Ca²⁺ to the mitochondrial Ca²⁺ uniporter (MCU), thereby shaping cytosolic Ca²⁺ signals and regulating mitochondrial function and cell fate (Kirichok et al., 2004; Mendes et al., 2005; De Stefani et al., 2011; Rowland and Voeltz, 2012; Bartok et al., 2019; Vecellio Reane et al., 2020). Recent observations suggest that all three IP₃R subtypes (IP₃R1–3) localize to ERMCSs and support Ca²⁺ delivery to mitochondria (Bartok et al., 2019). ERMCSs are held together by tethers (Vecellio Reane et al., 2020), with tripartite interactions among IP₃Rs, the anchoring cytosolic protein GRP75 (also known as HSPA9) and the voltage-dependent anion-selective channel protein 1 (VDAC1) in the outer mitochondrial membrane (OMM), most thoroughly implicated in regulating Ca²⁺ transfer (Gincel et al., 2001; Rapizzi et al., 2002; De Stefani et al., 2012). Disruption of ERMCSs is implicated in neurodegenerative diseases (Milakovic et al., 2006; Area-Gomez et al., 2009; Kim et al., 2010; Yu et al., 2021), metabolic disorders (Tubbs and Rieusset, 2017) and cancer (Simoes et al., 2020).

Only a fraction of the IP₃Rs expressed in a cell (∼30% in HeLa cells) release Ca²⁺ into the cytosol in response to IP₃; these have been described as ‘licensed’ IP₃Rs (Thillaiappan et al., 2017). Evidence that Kras-induced actin-binding protein (KRAP; also known as IP₃ receptor-interacting domain-containing protein 2 or ITPRID2) associates with IP₃Rs and actin (Fujimoto et al., 2011a,b) and perhaps influences Ca²⁺ release (Fujimoto et al., 2011c) prompted further analysis of its role in licensing IP₃Rs. KRAP is ubiquitously expressed, its amino acid sequence is well-conserved across species (Inokuchi et al., 2004; Fujimoto et al., 2007; Fujimoto and Shirasawa, 2011), its interaction with IP₃R requires a double phenylalanine motif within the N-terminal region of KRAP (Fujimoto et al., 2011b), and loss of KRAP alters the subcellular distribution of IP₃Rs (Fujimoto et al., 2011a; Thillaiappan et al., 2021). It is now clear that licensed IP₃Rs are tethered in small clusters by KRAP to actin near the plasma membrane and adjacent to the sites where store-operated Ca²⁺ entry occurs (Thillaiappan et al., 2021; Vorontsova et al., 2022).

Loss of KRAP abolishes the cytosolic Ca²⁺ signals evoked by IP₃Rs, which are initiated by KRAP-licensed IP₃Rs parked immediately beneath the plasma membrane (PM). The functions of the many IP₃Rs that are neither immobilized nor associated with KRAP are unresolved (Thillaiappan et al., 2021). It is not known, for example, whether the IP₃Rs that transfer Ca²⁺ at MCSs to the cytosolic surface of organelles are, similar to the IP₃Rs that evoke cytosolic Ca²⁺ signals, licensed by KRAP. As the positioning of IP₃Rs within specialized ERMCSs is critical for transfer of Ca²⁺ via MCU (Kirichok et al., 2004; Mendes et al., 2005; De Stefani et al., 2011; Rowland Voeltz, 2012; Bartok et al., 2019; Vecellio Reane et al., 2020), we assessed whether KRAP also regulates the IP₃Rs that transport Ca²⁺ at these MCSs. The role of KRAP at ERMCSs has not hitherto been studied.

Here, we show that KRAP both regulates the activity of the IP₃Rs that deliver Ca²⁺ to mitochondria and structurally stabilizes the ERMCSs. We conclude that KRAP exerts a dual regulation of Ca²⁺ transfer from ER to mitochondria.

RESULTS

IP₃-evoked cytosolic and mitochondrial Ca²⁺ signals require KRAP

We used validated Ca²⁺ indicators to simultaneously record the free Ca²⁺ concentrations in the cytosol ([Ca²⁺]_c) and mitochondrial matrix ([Ca²⁺]_m) of single HeLa cells (Fig. 1A; Fig. S1; Materials and Methods). Stimulation of HeLa cells in Ca²⁺-free HEPES-buffered saline (HBS) with a maximally effective concentration of histamine to evoke IP₃ formation through endogenous receptors caused a rapid increase in both [Ca²⁺]_c and [Ca²⁺]_m (Fig. 1B). Pre-treatment of cells with carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) to dissipate the H⁺ gradient across the inner mitochondrial membrane had no effect on the histamine-evoked increase in [Ca²⁺]_c, but it almost abolished the increase in [Ca²⁺]_m (Fig. 1C). We used siRNA to reduce expression of MCU by 78±7% (mean±s.e.m.) (Fig. 1D). This treatment almost abolished the histamine-evoked increase in [Ca²⁺]_m without significantly affecting the increase in [Ca²⁺]_c (Fig. 1E–G). These results, which are consistent with many published observations (Rapizzi et al., 2002; De Stefani et al., 2011, 2012; Bartok et al., 2019; Katona et al., 2022), confirm that IP₃-evoked Ca²⁺ release causes rapid uptake of Ca²⁺ into mitochondria via MCU.

Fig. 1.

Histamine evokes transfer of Ca²⁺ from the ER to mitochondria through IP₃Rs. (A) Wide-field fluorescence images of HeLa cells expressing TOM20–GFP (green) and Mito-R-GECO1 (pseudo-coloured in magenta); the overlay shows colocalization in white. Images are typical of four independent experiments. Manders' coefficients for the fraction of Mito-R-GECO1 colocalizing with TOM20–GFP=0.88±0.07 (mean±s.d.), n=4. (B) HeLa cells expressing Mito-R-GECO1 (mitochondrial Ca²⁺ indicator) and loaded with Calbryte 520 (cytosolic Ca²⁺ indicator) were stimulated with histamine (100 µM) in Ca²⁺-free HBS. Fluorescence from the Ca²⁺-saturated indicators (F_max) was determined by addition of CGP 37157 (CGP, 50 µM) to inhibit the mitochondrial Na⁺/Ca²⁺ exchanger and then ionomycin (Iono, 10 μM) with 20 mM CaCl₂. Results (F/F_max) show fluorescence recorded from an entire cell. Results are typical of at least ten experiments. (C) Summary results show peak changes in fluorescence (ΔF/F_max) evoked by histamine (100 µM) alone, or after pre-incubation with FCCP (5 µM, 10 min). Results show values for individual cells and mean±s.d. from ten (control) or 15 cells (FCCP) from three independent experiments. ns, not significant, P>0.05; **P<0.01; one-way ANOVA with Tukey's correction. (D) Western blots showing the effect of non-silencing (NS) siRNA or siRNA directed at MCU in HeLa cells. Protein loadings (μg/lane) and molecular mass markers (kDa) are shown. Results are typical of three independent experiments. MCU siRNA reduced MCU expression to 22±7% of its expression in cells treated with NS siRNA (mean±s.e.m., n=3). (E,F) Effects of histamine (100 µM) in Ca²⁺-free HBS on [Ca²⁺]_c (Calbryte 520) (E) or [Ca²⁺]_m (Mito-R-GECO1) (F) in single HeLa cells treated with NS or MCU siRNA. Results, reported as F/F_max, are typical of 15–16 cells from three independent experiments. (G) Summary results show effects of histamine in HeLa cells treated with NS or MCU siRNA. Results (individual values, mean±s.d.) are from 16 (NS) or 15 cells (MCU siRNA) from three independent experiments. ns, not significant, P>0.05; **P<0.01; one-way ANOVA with Tukey’s correction.

Treatment of HeLa cells with an appropriate siRNA caused a substantial (83±9%, mean±s.e.m.) reduction in KRAP expression (Fig. 2A) and abolished the histamine-evoked increases in both [Ca²⁺]_c and [Ca²⁺]_m. Both responses were rescued by expression of a siRNA-resistant KRAP (Fig. 2B–E). Furthermore, overexpression of KRAP in HeLa cells (352±32% of control levels, mean±s.e.m.) exaggerated the histamine-evoked increases in both [Ca²⁺]_c and [Ca²⁺]_m (Fig. 3A–E). We conclude that both the global cytosolic Ca²⁺ signals evoked by IP₃ and the uptake of Ca²⁺ by mitochondria through MCU require KRAP.

Fig. 2.

KRAP is required for transfer of Ca²⁺ from the ER to mitochondria through IP₃R. (A) Western blots from HeLa cells showing the effect of NS siRNA or siRNA directed at KRAP alone or after expression of siRNA-resistant KRAP. Results are typical of four experiments. Protein loadings (μg/lane) and molecular mass markers (kDa, left) are shown. (B,C) Effects of histamine (100 µM) in Ca²⁺-free HBS on [Ca²⁺]_c (Calbryte 520) (B) or [Ca²⁺]_m (Mito-R-GECO1) (C) in single HeLa cells treated with NS siRNA, KRAP siRNA or KRAP siRNA with siRNA-resistant KRAP (‘rescue’). Results, reported as F/F_max, are typical of 14–23 cells from three independent experiments. (D,E) Summary results show peak changes in fluorescence (ΔF/F_max) from cytosolic (D) or mitochondrial (E) Ca²⁺ indicators. Results show individual values with mean±s.d. from 14 (NS siRNA), 23 (KRAP siRNA) and 23 cells (rescue) from three independent experiments. ***P<0.001; ****P<0.0001; one-way ANOVA with Tukey’s correction.

Fig. 3.

KRAP regulates IP₃R-mediated transfer of Ca²⁺ from the ER to mitochondria without affecting MCU. (A) Western blots from control HeLa cells and after overexpression of KRAP. KRAP expression was 3.5±0.3-fold greater in cells overexpressing KRAP (mean±s.e.m., three independent analyses). Protein loadings (μg/lane) and molecular mass markers (kDa) are shown. (B,C) Effects of histamine (100 µM) in Ca²⁺-free HBS on [Ca²⁺]_c (Calbryte 520) (B) or [Ca²⁺]_m (Mito-R-GECO1) (C) in single HeLa cells overexpressing KRAP or after mock transfection with an empty vector. (D,E) Summary results show peak changes in cytosolic (D) and mitochondrial (E) fluorescence from Ca²⁺ indicators evoked by histamine (100 µM) in HeLa cells overexpressing KRAP or after mock transfection. Individual values with mean±s.d. are shown from six (mock) or 12 (KRAP) cells. Cells were transfected with plasmids expressing Mito-R-GECO and KRAP (in a 1:3 ratio), and all cells expressing Mito-R-GECO in the field were included in the analysis. *P<0.05; two-tailed unpaired Student's t-test. (F) Western blots from HeLa cells showing the effect of NS siRNA or siRNA directed against KRAP on the expression of MCU. KRAP expression was reduced by 70±6% (mean±s.e.m.). MCU expression was 97±24% of its level in NS siRNA-treated cells. P>0.05; two-tailed paired Student's t-test of MCU expression in KRAP siRNA-treated cells compared to that in NS siRNA-treated cells from six independent experiments. (G) HeLa cells expressing Mito-R-GECO1 and loaded with Calbryte 520 were stimulated with thapsigargin (TG, 5 µM) in Ca²⁺-free HBS, and then CGP 37157 (CGP, 50 µM) and ionomycin (Iono, 10 μM) with 20 mM CaCl₂ to obtain F_max. Results (F/F_max) show fluorescence recorded from an entire cell for each indicator. Results are typical of at least three experiments. (H) Summary results show peak changes in fluorescence (ΔF/F_max) from cytosolic or mitochondrial Ca²⁺ indicators evoked by thapsigargin (5 µM). Mean±s.d. (and individual values) are shown from 20 (NS siRNA) and 15 (KRAP siRNA) cells from three independent experiments. ns, not significant, P>0.05; one-way ANOVA with Tukey’s correction.

KRAP is required for delivery of Ca²⁺ from IP₃R to mitochondria

We next considered whether the requirement for KRAP for mitochondrial Ca²⁺ uptake arises from KRAP regulating delivery of Ca²⁺ to the OMM and/or regulation of MCU. Treatment of HeLa cells with KRAP siRNA had no effect on MCU expression [MCU expression was 97±24% of its level in non-silencing (NS) siRNA-treated cells, mean±s.e.m., n=6] (Fig. 3F) or mitochondrial membrane potential (Fig. S2). HeLa cells were stimulated with thapsigargin in Ca²⁺-free HBS to promote loss of Ca²⁺ from the ER, which is independent of IP₃R activity. Thapsigargin evoked increases in both [Ca²⁺]_c and [Ca²⁺]_m, but neither response was affected by siRNA-mediated loss of KRAP expression (Fig. 3G,H).

These results establish that KRAP is required for mitochondria to accumulate Ca²⁺ released through IP₃Rs. Furthermore, the deficient Ca²⁺ transfer in the absence of KRAP is not due to collapse of the mitochondrial membrane potential, loss of MCU expression or compromised function of MCU.

KRAP mediates association of IP₃R with mitochondria

We used HeLa cells in which endogenous IP₃R1 (encoded by ITPR1) was tagged with EGFP (EGFP–IP₃R1 HeLa cells), an antibody to KRAP and MitoTracker to explore the relationships among IP₃Rs, mitochondria and KRAP. Although only IP₃R1 is visible in EGFP–IP₃R1 HeLa cells, it is the major IP₃R subtype in these cells (Itzhak et al., 2016) and it assembles into tetrameric channels with the other subtypes (Thillaiappan et al., 2017). We assume, therefore, that EGFP identifies most IP₃Rs in EGFP–IP₃R1 HeLa cells (Thillaiappan et al., 2021). Confocal microscopy established that 24±11% of EGFP–IP₃Rs (mean±s.d., n=10 cells) were associated with mitochondria and many of these IP₃R puncta were associated with KRAP (Fig. 4A,B). Quantitative analyses using an object-based method (Gilles et al., 2017) established that the colocalization of KRAP with IP₃Rs was indistinguishable for IP₃Rs associated with mitochondria (32±13%, mean±s.d., n=10) and for IP₃Rs with no mitochondrial association (28±9%, n=10) (Fig. 4C). Analysis of nearest-neighbour distances revealed that IP₃Rs affiliated with mitochondria were significantly more colocalized with KRAP compared to randomly distributed IP₃R puncta (Fig. 4D). These results establish that a significant fraction of the IP₃Rs that associate with mitochondria are also associated with KRAP. Our conclusion is consistent with a proteomic analysis that identified KRAP near the OMM (Hung et al., 2017).

Fig. 4.

Colocalization of KRAP and IP₃R with mitochondria. (A) Confocal section near the plasma membrane of an EGFP–IP₃R1 HeLa cell loaded with MitoTracker Deep Red (pseudo-coloured in blue) and immunostained for KRAP (magenta). GFP Booster was used to enhance the fluorescence intensity of the EGFP signal. Images show some IP₃R puncta (green) colocalized with KRAP (white where colocalized) associated with mitochondria. Boxed regions are enlarged below. Images are typical of ten cells. (B) Fluorescence intensity profiles of EGFP–IP₃R1, KRAP and mitochondria from the regions marked by white lines and labelled i–iii in A. FU, fluorescence units. (C) Percentage of mitochondria-associated IP₃R puncta or non-mitochondria-associated IP₃R puncta that colocalize with KRAP (centre-to-centre separations <233 nm). Results show individual values with mean±s.d. for ten cells. P>0.05; two-tailed paired Student's t-test. (D) Frequency distributions of centre-to-centre distances for mitochondria-associated IP₃R puncta (2071 puncta from ten cells) and the nearest KRAP punctum. Observed values and values after randomization of positions of the KRAP puncta (100 iterations) are shown. 31.7±12.9% (mean±s.d., n=10 cells) of IP₃R puncta are within 233 nm (red line) of a KRAP punctum (5.2±2.7% after randomization; P<0.0001, two-tailed paired Student's t-test). Images and results are representative of ten independent dishes.

We used proximity ligation assays (PLAs), which detect proteins closer than about 40 nm (Tubbs and Rieusset, 2016), to examine the association of VDAC1, a protein embedded in the OMM, with IP₃Rs and KRAP. The results demonstrate that VDAC1 is in close proximity to both IP₃Rs and KRAP in EGFP–IP₃R1 HeLa cells (Fig. 5A,B). Furthermore, EGFP–IP₃R puncta and the PLA spots indicative of KRAP and VDAC1 proximity were significantly colocalized (Manders' split coefficient, 0.82±0.15; n=7 cells, Costes' P-value, 100%) (Fig. 5C). These results confirm that many IP₃Rs that are affiliated with mitochondria are closely associated with KRAP.

Fig. 5.

KRAP contributes to the association of IP₃R with mitochondria. (A) Proximity ligation assay (PLA) uses primary antibodies that recognize proteins in the ER or mitochondrial membranes. Hybridization of complementary oligonucleotides conjugated to secondary antibodies occurs if the primary antibodies are within ∼40 nm of each other. Ligation of the hybridized strands then allows rolling circle amplification (RCA) of the oligonucleotide and incorporation of the red fluorescent probe. PLA analyses of EGFP–IP₃R1 HeLa cells used combinations of antibodies (Ab) to VDAC1, KRAP or GFP (for EGFP–IP₃R1). Maximum-intensity z-projections of confocal images show the PLA dots (red) and nuclei (DAPI, blue). Control images with only single antibodies are also shown. Results are typical of three to five independent PLA analyses. (B) Summary results show the average number of PLA dots per cell in a field that typically included ∼10 cells. Results show values for each field with mean±s.e.m. For each analysis, the number of cells, fields and independent experiments were: GFP and VDAC1 (512, 23, 4), KRAP and VDAC1 (983, 37, 6), KRAP only (559, 23, 3), GFP only (767, 26, 3) and VDAC1 only (738, 34, 4). *P<0.05; ****P<0.0001; one-way ANOVA with Tukey’s correction. (C) Confocal section of PLA in EGFP–IP₃R1 HeLa cells shows KRAP proximity to VDAC1 (PLA dots, magenta) and endogenously tagged IP₃R1 (green). Enlargements of boxed areas show coincidence of EGFP–IP₃R puncta with PLA spots. Manders' split coefficient, 0.82±0.15 (mean±s.d.); n=7 cells. (D) PLA analyses of the interactions between VDAC1 and EGFP–IP₃R1 in EGFP–IP₃R1 HeLa cells treated with NS siRNA or KRAP siRNA and after rescue with an siRNA-resistant KRAP expression plasmid. Maximum-intensity z-projections of confocal images show the PLA dots (red) and nuclei (DAPI, blue). (E) Summary results show the average number of PLA dots per cell in a field, with mean±s.e.m. For each treatment, the number of cells, fields and independent experiments were: NS siRNA (1116, 46, 7), KRAP siRNA (1003, 44, 7) and rescue (753, 51, 7). *P<0.05; repeated measures one-way ANOVA with Bonferroni's test performed on the seven matched experiments. (F) IP₃Rs at ER–mitochondrial membrane contact sites (ERMCSs) release Ca²⁺ that fuels the low-affinity mitochondrial Ca²⁺ uptake system (MCU). KRAP licenses these IP₃Rs to release Ca²⁺ and it also stabilizes the ERMCS.

Next, we used PLA with EGFP–IP₃R1 and VDAC1 to determine the effects of KRAP on the association of IP₃Rs with mitochondria. siRNA-mediated knockdown of KRAP caused a significant reduction in the PLA signal. The association of IP₃Rs with mitochondria was restored by the expression of siRNA-resistant KRAP (Fig. 5D,E). These results establish that KRAP contributes to maintaining the ERMCSs wherein IP₃Rs and mitochondria interact to facilitate the delivery of Ca²⁺ from the ER to mitochondria (Fig. 5F).

DISCUSSION

We showed previously that IP₃Rs must be licensed by their association with KRAP before they can respond to IP₃ and evoke local or global cytosolic Ca²⁺ signals (Thillaiappan et al., 2021). Our present results extend these observations by demonstrating that KRAP is also required to license the IP₃Rs that deliver Ca²⁺ to the surface of mitochondria within ERMCSs (Figs 1–3). KRAP is present at ERMCSs, a substantial fraction of mitochondria-associated IP₃Rs (∼30%) colocalize with KRAP, and loss of KRAP abolishes the mitochondrial Ca²⁺ uptake that usually follows activation of IP₃Rs (Figs 1–5). As the loss of mitochondrial Ca²⁺ uptake is not due to collapse of the mitochondrial membrane potential (Fig. S2) or deficient MCU (Fig. 3G,H), we suggest that regardless of whether IP₃Rs deliver Ca²⁺ to the cytosol or within ERMCSs, their activity must be licensed by association with KRAP (Fig. 5F). Within both the cytosol and ERMCSs, only ∼30% of IP₃Rs are associated with KRAP, suggesting that, in both locations, most IP₃Rs are incapable of mediating Ca²⁺ release.

We note in passing that dissipating the mitochondrial membrane potential (with FCCP) or inhibition of MCU expression abolished IP₃-evoked mitochondrial Ca²⁺ uptake, but there was no associated ‘overflow’ increase in [Ca²⁺]_c (Fig. 1C–G). Similar results have been reported by others (Young et al., 2022), although, more commonly, inhibition of mitochondrial Ca²⁺ uptake has been reported to exaggerate the IP₃-evoked increase in [Ca²⁺]_c (Drago et al., 2012; Fonteriz et al., 2016; Koval et al., 2019; Lombardi et al., 2019). The disparity probably reflects the differential distribution of licensed IP₃Rs between ERMCSs and cytosol-disposed ER in different cells; only cells with many of their licensed IP₃Rs within ERMCSs are likely to show detectable ‘overflow’ increases in [Ca²⁺]_c when mitochondrial Ca²⁺ uptake is inhibited.

Our results identified an additional role for KRAP in facilitating Ca²⁺ transfer at ERMCSs, namely in maintaining the assembly of the ERMCSs wherein Ca²⁺ transfer from IP₃Rs to mitochondria occurs. PLA analyses of KRAP with VDAC1 (in the OMM) showed that they were closely associated (within ∼40 nm of each other) at ERMCSs, and similar analyses with VDAC1 and IP₃Rs established that their association was disrupted when KRAP expression was reduced (Fig. 5). The PLA analyses were restricted to endogenously tagged EGFP–IP₃R1 because the GFP antibody is better than any available IP₃R antibodies. However, our results are likely to be applicable to all three IP₃R subtypes. First, within the HeLa cells used for PLA, the association of EGFP–IP₃R1 with the other subtypes is consistent with their random assembly into tetramers (Thillaiappan et al., 2017). Second, recent work shows that all three IP₃R subtypes contribute to Ca²⁺ transfer between the ER and mitochondria (Bartok et al., 2019; Katona et al., 2022). We have not therefore assessed whether KRAP differentially regulates IP₃R subtypes at ERMCSs.

We suggest that KRAP is required for effective assembly or for maintaining the stability of ERMCSs wherein IP₃R delivers Ca²⁺ to mitochondria (Fig. 5F). This observation aligns with a developing theme in Ca²⁺ signalling, namely that the Ca²⁺ channels responsible for delivering Ca²⁺ into MCSs might independently contribute to the structural stability of the MCSs. IP₃Rs, for example, have been reported to contribute to the stability of ERMCSs (Bartok et al., 2019), and the endosomal Ca²⁺ channel TPC1 (also known as TPCN1) has been reported to stabilize the MCS between the ER and endosomes (Kilpatrick et al., 2017). Within ERMCSs, the structural contribution of IP₃Rs to MCS stability occurred even with pore-dead Ca²⁺ channels (Bartok et al., 2019; Katona et al., 2022). Hence, the ∼70% of IP₃Rs at ERMCSs that are not associated with KRAP and thus incapable of releasing Ca²⁺ might nevertheless contribute to assembling the ERMCSs. Our results develop this theme further by revealing that KRAP, a protein essential for IP₃R activity, also contributes to the structural stability of ERMCSs. IP₃Rs can move in and out of ERMCSs, but they must be immobilized to effectively transfer Ca²⁺ to mitochondria (Katona et al., 2022). Our results suggest that KRAP, in addition to its role in licensing IP₃Rs beneath the PM (Thillaiappan et al., 2017, 2021), also immobilizes and licenses IP₃Rs deeper within the cell at ERMCSs. We have not resolved whether the enhanced stability of ERMCSs by KRAP is mediated by immobilized IP₃Rs or by other interactions between KRAP and the ER and/or mitochondrial proteins.

A protein related to KRAP, TESPA1 (thymocyte-expressed, positive selection-associated 1), which shares an IP₃R-binding motif with KRAP, also interacts with IP₃Rs at ERMCSs and regulates cytosolic and mitochondrial Ca²⁺ signals (Matsuzaki et al., 2013). However, unlike the ubiquitously expressed KRAP (Fujimoto et al., 2007), TESPA1 is expressed predominantly in immune cells (Matsuzaki et al., 2012), such that a widespread regulation of ERMCSs by TESPA1 is unlikely.

We conclude that KRAP, originally identified because it is upregulated in colorectal cancer cell lines expressing mutant KRas (Inokuchi et al., 2004), fulfils a dual function in regulating Ca²⁺ transfer from the ER to mitochondria through IP₃Rs. KRAP licenses IP₃R within ERMCSs to respond to IP₃, just as it does for IP₃Rs that release Ca²⁺ to the cytosol (Thillaiappan et al., 2021). In addition, KRAP contributes to the structural stability of the ERMCSs wherein the Ca²⁺ transfer from the ER to mitochondria occurs. Remodelling of mitochondrial Ca²⁺ signalling is critical in shaping Ras-dependent oncogenic signalling cascades (Rimessi et al., 2015). Therefore, our findings also link the regulation of IP₃R-mediated Ca²⁺ delivery at ERMCSs to Ras signalling.

MATERIALS AND METHODS

Materials

Calbryte 520 AM was from AAT Bioquest (Sunnyvale, CA, USA). Histamine dihydrochloride, Triton X-100, FCCP, ECL western blotting detection reagents (Cytiva) and poly-L-lysine (0.1% w/v in H₂O) were from Sigma-Aldrich. Bovine serum albumin (BSA) was from Europa Bio-Products (Ely, UK). Human fibronectin was from Merck Millipore (Watford, UK). CGP 37157 [7-chloro-5-(2-chlorophenyl)-1,5-dihydro-4,1-benzothiazepin-2(3H)-one] and ionomycin were from Cambridge Biosciences (Cambridge, UK). MitoTracker Deep Red FM, NuPAGE 4–12% Bis–Tris gels (NB0322BOX), iBlot 2 transfer stacks, polyvinyl difluoride (PVDF) regular size membranes (IB24001), iBlot gel transfer system, PBS and the Neon transfection system were from Thermo Fisher Scientific. TransIT-LT1 reagent was from Geneflow (Lichfield, UK). Imaging dishes were from Cellvis (Gerasdorf bei Wein, Austria). Plasmids encoding the following proteins were used: TOM20 [translocase of outer (mitochondrial) membrane 20, also known as TOMM20] tagged to GFP (TOM20–GFP) (Prole and Taylor, 2019a), Mito-R-GECO1 (Addgene, #46021) (Wu et al., 2013), human KRAP with an N-terminal myc-DDK tag (OriGene, Rockville, MD, USA, #RC205550) and siRNA-resistant human KRAP with an N-terminal myc-DDK tag (Thillaiappan et al., 2021). The following antibodies were used for western blotting (WB), immunocytochemistry (IC), or PLAs: anti-KRAP (rabbit polyclonal; ProteinTech, Manchester, UK, #14157-1-AP, RRID:AB_2195472; WB 1:1000), ChromoTek GFP-Booster ATTO488 (nanobody; Chromotek, Planegg-Martinsried, Germany #gba488; RRID: AB_2631386; IC 1:500), anti-MCU (rabbit polyclonal; Thermo Fisher Scientific, PA5-109304, RRID:AB_2854715; WB 1:1000), anti-β-actin (mouse monoclonal; Cell Signaling Technology, Leiden, Netherlands, #8H10D10, RRID:AB_2242334; WB 1:20,000), IRDye 680RD goat anti-mouse IgG secondary antibody (LI-COR, 926-68070, RRID:AB_10956588; WB 1:10,000), horseradish peroxidase (HRP)-conjugated sheep anti-rabbit IgG H&L (Abcam, ab6795, RRID:AB_955446; WB 1:10,000), Alexa Fluor 568-conjugated goat anti-rabbit IgG (H+L) (Thermo Fisher Scientific, A-11011, RRID:AB_143157; WB 1:10,000), Alexa Fluor 488-conjugated goat anti-rabbit IgG (H+L) (Thermo Fisher Scientific, A-11008, RRID:AB_143165; WB 1:10,000), Alexa Fluor 568-conjugated goat anti-mouse IgG (H+L) (Thermo Fisher Scientific, A-21235, RRID:AB_2535804; WB 1:10,000), anti-VDAC1 (Abcam, ab14734, RRID:AB_443084; PLA 1:400), anti-GFP (Abcam, ab290, RRID:AB_303395; PLA 1:200, WB 1:1000) and anti-TOM20 (Abcam, ab56783, RRID:AB_945896; WB 1:1000). The PLA kits used were Duolink in situ PLA probes (anti-rabbit PLUS and anti-mouse MINUS, affinity-purified donkey anti-rabbit and anti-mouse IgG, respectively), Duolink in situ detection reagents Red, mounting medium containing DAPI and wash buffer, from Sigma-Aldrich or NaveniFlex 100 MR (from Cambridge Biosciences). The siRNAs used were: Silencer siRNA against human KRAP (Thermo Fisher Scientific, #AM16708, assay ID, 143004), Silencer negative control no.1 (Thermo Fisher Scientific, #AM4635) and ON-TARGETplusTM SMARTpool MCU siRNA (Dharmacon, L-015519-02-0005). All siRNA sequences are provided in Table S1. The Bradford protein assay kit (DC protein assay kit II, 5000112) and MOPS SDS Running Buffer 20× (NP001-02) were from Bio-Rad. Tetramethylrhodamine ethyl ester (TMRE)-mitochondrial membrane potential assay kit was from Abcam (ab113852). The sources of additional materials are provided in relevant sections of the Materials and Methods.

Cell culture and transfection

HeLa and edited HeLa cells, in which both copies of the endogenous genes encoding IP₃R1 (ITPR1) were N-terminally tagged with monomeric EGFP (EGFP–IP₃R1 HeLa cells) (Thillaiappan et al., 2017), were cultured in Dulbecco's modified Eagle medium (DMEM)/F-12 with GlutaMAX (Thermo Fisher Scientific, #31331028) and 10% foetal bovine serum (FBS; Sigma-Aldrich, #F7524, batch BCBX5042). The cells were maintained at 37°C in humidified air with 5% CO₂ and passaged every 3–4 days with TrypLE Express (Thermo Fisher Scientific, #12605010). Regular screening confirmed that all cells were mycoplasma free.

For imaging, cells were grown on 35-mm glass-bottomed dishes coated with human fibronectin (10 μg/ml). For transient transfections with protein-encoding plasmids, cells were transfected using TransIT-LT1 reagent (1.5 μg DNA per 3.75 μl reagent per dish) according to the manufacturer's instructions. Cells were assayed after 24 h. Transfections with siRNA followed the manufacturer's instructions using a Neon transfection system (Thermo Fisher Scientific). Briefly, ∼100,000 cells were diluted in 100 μl of R buffer and incubated with siRNA (200 nM). Cells were electroporated using a 100-μl Neon pipette tip (pulse voltage 1005 V; pulse width, 35 ms; two pulses) and transferred into complete growth medium. Electroporated cells were distributed into fibronectin-coated 35-mm imaging dishes (∼50,000 cells/dish) for microscopy or into 25-cm² culture flasks (∼100,000 cells) for protein analyses. Cells were assayed after 72 h. For rescue experiments, cells were first electroporated with siRNA (200 nM), and after 48 h, they were transfected with an siRNA-resistant KRAP plasmid (4.5 μg) using TransIT-LT1 reagent. Cells were used after a further 24 h.

Western blotting

Cells in 25-cm² culture flasks were harvested by centrifugation (650 g, 2 min, 22°C), washed with ice-cold PBS (1.06 mM KH₂PO₄, 155 mM NaCl, 3 mM Na₂HPO₄, pH 7.3, 20°C), and the pellet was resuspended (0.1 ml) in radioimmunoprecipitation assay buffer (RIPA; 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 8.0) containing protease inhibitors (Sigma-Aldrich, #11873580001; 1 mini-tablet per 10 ml). After sonication on ice (Transsonic, T420; three 10 s sonication pulses and three 10 s pauses), samples were incubated with rotation (30 min, 4°C) and centrifuged (14,000 g, 15 min, 4°C), and the supernatant was used for analysis. Proteins were quantified using the DC protein assay kit II with BSA in RIPA buffer as the standard. For western blotting, proteins were separated on NuPAGE 4–12% Bis-Tris gels, run at 160 V for 45 min in MOPS SDS running buffer, and transferred to a PVDF membrane using an iBlot 2 gel-transfer system. The membrane was blocked (1 h with shaking) in TBS (137 mM NaCl, 20 mM Tris, pH 7.6) containing 0.1% Tween 20 (TBST) with 5% BSA, washed twice with TBST, incubated in TBST containing 1% BSA (16 h, 4°C with shaking) with the primary antibody, washed with TBST (five times for 5 min), incubated with secondary antibody in TBST containing 1% BSA (1 h, 20°C with shaking), and washed with TBST (five times for 5 min). Bands were quantified using an iBright FL1500 imaging system (Thermo Fisher Scientific) either directly for the fluorescent secondary antibody (Alexa Fluor 568) or by chemiluminescence after treatment with ECL Prime western blotting detection reagent (for HRP-conjugated secondary antibodies; see Materials). We confirmed that, within the range of loadings used for analysis, protein band intensities scaled linearly with the amount of protein loaded.

Immunocytochemistry

Cells grown on fibronectin-coated 35-mm glass-bottomed dishes were washed three times with HBS (20°C) and then incubated with MitoTracker Deep Red (500 nM, 30 min, 37°C). HBS had the following composition: 135 mM NaCl, 5.9 mM KCl, 1.2 mM MgCl₂, 1.5 mM CaCl₂, 11.5 mM d-glucose, 11.6 mM HEPES, pH 7.3. Cells were washed three times with HBS, fixed with methanol (20 min, −20°C), washed three times with PBS, and permeabilized using 0.25% Triton X-100 in PBS (5 min, 20°C). After three washes with PBS, cells were incubated in PBS containing 5% BSA (1 h, 20°C), washed in the same medium, then incubated (12 h, 4°C) with ChromoTek GFP-Booster ATTO488 and the primary antibody to KRAP in PBS containing 3% BSA. Cells were then washed three times with PBS, incubated in the dark with the secondary antibody in PBS with 3% BSA (1 h, 20°C), and washed three times with PBS before imaging. Dilutions of the antibodies used are provided in the Materials section.

Proximity ligation assays

A Duolink PLA was used to determine protein interactions at ERMCSs. PLA reports the presence of juxtaposed proteins when secondary antibodies conjugated to complementary oligonucleotides come sufficiently close (∼40 nm) to hybridize and allow rolling-circle amplification, which is then detected with a fluorescent probe (Texas Red) (Tubbs and Rieusset, 2016) (Fig. 5A). Cells grown on fibronectin-coated 35-mm glass-bottomed dishes were fixed in PBS with 4% paraformaldehyde (30 min, 20°C), washed with PBS, permeabilized (0.25% Triton X-100 in PBS, 5 min, 20°C) and incubated with primary antibodies (16 h, 4°C). All subsequent steps, including incubation with the oligonucleotide-conjugated secondary antibodies (anti-rabbit PLUS and anti-mouse MINUS), were performed according to the manufacturer's instructions. The antibodies and their dilutions are described in the Materials section. PLA products were visualized using an Olympus microscope equipped with a 60× objective, and the number of spots per cell were quantified using ImageJ. PLA analysis of interactions between EGFP–IP₃R and VDAC1 or KRAP and VDAC1 used Duolink PLA reagents (Fig. 5A,B). Supply difficulties necessitated the use of a different PLA kit (NaveniFlex) for analyses of KRAP knockdown (Fig. 5D,E). Indistinguishable results were obtained when the two kits were applied to analyses of the interactions between IP₃R on the ER and VDAC1 on the mitochondria.

Fluorescence microscopy

Fluorescence microscopy used an Olympus (Ca²⁺ imaging, colocalization of Mito-R-GECO and PLA analysis) or Zeiss (three-colour imaging of IP₃R, KRAP and mitochondria) microscope. The inverted Olympus IX83 microscope was equipped with 60× (numerical aperture or NA, 1.45) and 100× oil-immersion objectives (NA, 1.49), a multi-line laser bank (488 and 561 nm) and an iLas² targeted laser illumination system (Cairn, Faversham, UK). The excitation light passed through a quad dichroic beam-splitter (TRF89902-QUAD). The emitted light passed through appropriate filters (Cairn Optospin; peak/bandwidth (nm): 525/50 and 630/75) and was detected with an iXon Ultra 897 electron multiplied charge-coupled device (EMCCD) camera (512×512 pixels, Andor, Belfast, Northern Ireland). The Zeiss LSM700 upright confocal microscope, on which imaging dishes were viewed in an inverted position, was equipped with a 63× oil-immersion objective (NA, 1.4) and four excitation solid-state laser lines (405, 488, 555 and 639 nm). The emitted light was detected by two photomultiplier tubes in the scanhead using a shared pinhole and a sliding dichroic filter to separate the emitted wavelengths. We confirmed that for all multi-colour imaging, there was no bleed-through between channels.

Measurement of cytosolic and mitochondrial Ca²⁺ concentrations in single cells

We sought cytosolic and mitochondrial Ca²⁺ indicators with compatible spectra and with changes in fluorescence amplitude after stimulation with a maximally effective concentration of histamine (100 μM) that were usually less than 50% of the maximal fluorescence signal (F_max). F_max signals were determined after addition of CGP 37157 (50 µM) to inhibit the mitochondrial Na⁺/Ca²⁺ exchanger (Suzuki et al., 2014) and then, after 1 min, addition of ionomycin (10 μM) in HBS containing a final Ca²⁺ concentration of 20 mM. With these conditions satisfied (i.e. the fluorescence signal F usually less than 50% of F_max), it is reasonable to assume that changes in the fluorescence intensity of the Ca²⁺ indicators are approximately linearly related to changes in [Ca²⁺]. After investigating several cytosolic (Fluo 8 AM, Cal 520 AM and Calbryte 520 AM) and mitochondrial Ca²⁺ indicators (Fura 2 AM, MtCEPIA 2, Mito-LAR-GECO1.2 and Mito-R-GECO1), we selected Calbryte 520 (equilibrium dissociation constant for Ca²⁺, =1200 nM) and Mito-R-GECO1 (= 480 nM) (Wu et al., 2013) to measure [Ca²⁺]_c and [Ca²⁺]_m, respectively.

HeLa cells grown on fibronectin-coated 35-mm glass-bottomed dishes were transfected with Mito-R-GECO1 and then loaded with Calbryte 520 by incubation with Calbryte 520 AM (2 μM, 1 h, 20°C) in HBS supplemented with 0.02% Pluronic F-127 (Sigma-Aldrich, #P2443). Cells were then washed and incubated in HBS for 30 min, and, immediately before imaging, the HBS was replaced with Ca²⁺-free HBS (HBS with no added CaCl₂). Imaging at 20°C used an Olympus IX83 microscope with a 100× objective and switches between pairs of excitation (λ_ex) and emission (λ_em) wavelengths: Calbryte 520 (λ_ex=488 nm, λ_em=525 nm) and Mito-R-GECO1 (λ_ex=561 nm, λ_em=630 nm). Images were acquired at 1-s intervals for each indicator, corrected for background fluorescence, and analyzed using MetaMorph Microscopy Automation and Image Analysis software (Molecular Devices, San Jose, CA, USA). A region of interest was drawn around each cell successfully transfected with Mito-R-GECO1 and responses for both cytosolic and mitochondrial Ca²⁺ indicators were reported as F/F_max. We confirmed that under the conditions used for recording [Ca²⁺]_c and [Ca²⁺]_m, there was no bleed-through between the two channels (Fig. S1).

Measurement of mitochondrial membrane potential

HeLa cells treated with NS siRNA or siRNA against KRAP were grown on poly-L-lysine-coated (0.01% w/v) 35-mm imaging dishes. Mitochondrial membrane potential was measured using a TMRE assay following the manufacturer's instructions. Briefly, cells were incubated with TMRE (100 nM, 15 min) and DAPI at 37°C in humidified air with 5% CO₂. Where indicated, cells were incubated with FCCP (20 µM, 10 min) to dissipate the mitochondrial membrane potential before addition of TMRE. Cells were then washed twice with PBS, and fluorescence intensity was measured from single cells at 20°C in PBS using an EVOS M7000 cell imaging system (Thermo Fisher Scientific) equipped with a 60× oil-immersion objective (NA, 1.45). Cells were kept in the dark during all manipulations.

Quantification of fluorescence images

Fluorescence images were corrected for background by subtraction of fluorescence collected from a region outside the cell using MetaMorph software. Capture and processing of widefield images and confocal PLA images was performed using MetaMorph. Confocal images of IP₃R, KRAP and mitochondrial fluorescence were captured using Zen imaging software (Carl Zeiss Microscopy, Germany).

Colocalization analysis within widefield images (Mito-R-GECO1 and TOM20–GFP) used the JACoP plugin (Bolte and Cordelieres, 2006) in Fiji/ImageJ (Schindelin et al., 2012) to calculate the Manders' coefficients that report the fraction of Mito-R-GECO1 colocalized with TOM20–GFP. For colocalization analyses of confocal images, images were first deconvolved using SVI HuygensPro software (Scientific Volume Imaging B.V., VB Hilversum, the Netherlands) and then background corrected using MetaMorph. Colocalization used an object-based program within ImageJ (DiAna) (Gilles et al., 2017), wherein the centre-to-centre distances between each EGFP–IP₃R punctum and the nearest KRAP punctum were assessed following a spot segmentation of each image. For analysis of three-colour colocalization (IP₃Rs, KRAP and mitochondria), a mask of the regions containing mitochondria was generated, EGFP–IP₃R puncta within these regions were identified and the centre-to-centre distances from these IP₃R puncta to the nearest KRAP punctum were assessed using DiAna. We considered IP₃R and KRAP puncta to be colocalized if their centre-to-centre distance was <233 nm (3 pixels), which is close to the theoretical lateral resolution of the confocal microscope.

PLA spots were quantified according to previously described methods (López-Cano et al., 2019). Briefly, following background subtraction, maximum z-projected images were converted into binary images. The number of PLA spots and cells in a field identified by DAPI staining were counted using the Analyze Particles function on ImageJ.

Statistical analysis

Most results are presented as mean±s.d. or s.e.m. from n independent analyses. Statistical analyses used either two-tailed paired or unpaired Student's t-test (two variables) or one-way ANOVA with Tukey's multiple comparisons test (more than two variables); P<0.05 was considered significant (PRISM, version 9, GraphPad, USA). For colocalization analyses using DiAna, the distribution of KRAP puncta was shuffled 100 times within each cell and the distances from each KRAP punctum to the nearest IP₃R punctum was reassessed. We considered the observed distances to be statistically significant if they fell outside the 95% confidence interval of the randomized distances.